Show HN: You can now train your own LLMs with less memory, on your consumer GPUs
0 comments
- Points4
- Comments0
#25Rank now 4
#25Peak rank
1Days on list
17Score · 7d
Nov 10, 2026First seen
Rank · 7 days
On Hacker News
*TL;DR: You can train your own 27B model on a 3090 with 24GB VRAM, though you would have to go through hundreds of billions of tokens for it to be good. I am working on conversion of existing models. Check repos for code and the paper if you are interested in more. I have a 40.4GB Qwen Flash* N*ext GGUF* Hi, I do AI research. I have been working on an optimizer that doesn't need the \~16 bytes of VRAM per weight to train a model that AdamW does. I focused on training ternary networks bcs…
- Points
- 4
- Comments
- 0
- Submitted by
- gladia-solaria
- Posted
- Nov 10, 2026
Story details from Hacker News ↗.
Trending history
Show HN: You can now train your own LLMs with less memory, on your consumer GPUs was first seen on Hacker News · Show HN on Nov 10, 2026. Since then it has appeared in 9 checks on 1 different day. Its best position was #25 of 30, reached on Nov 10, 18:21 UTC. It is #25 right now, climbing from #27 to #25 over the last 6 checks.
Rank history
Position on Hacker News · Show HN at every check (lower is better).
Last 14 days
Best rank per day (UTC).
25
Oct 28today
Metric history
points as reported by Hacker News.
Featured in
Archived daily, weekly and monthly lists that include this entry.
Show HN 1 day, best #27
Questions
Is Show HN: You can now train your own LLMs with less memory, on your consumer GPUs trending on Show HN right now?
Yes. As of Nov 10, 18:21 UTC it is #25, climbing from #27 to #25 over the last 6 checks.
What is the highest rank Show HN: You can now train your own LLMs with less memory, on your consumer GPUs reached?
#25 on Hacker News · Show HN, on Nov 10, 18:21 UTC. Trendoday checks this list every 30 minutes.
How long has Show HN: You can now train your own LLMs with less memory, on your consumer GPUs been trending?
It was first seen on Nov 10, 2026 and has been on the list on 1 day (9 checks) since.
How much discussion did it get on Hacker News?
4 points and 0 comments; submitted by gladia-solaria on Nov 10, 2026.